|
Abstract :
|
As artificial intelligence evolves from providing information and recommendations to acting on behalf of consumers, consumer protection in ICT services increasingly extends beyond personal-data processing to the control of delegated execution. AI agents may interpret user intent, invoke data and third-party services, and perform actions across terminals and platforms, creating new requirements for transparency, authorization, intervention and redress. This contribution reviews a layered set of practices emerging in China, including legal safeguards for personal information and automated decision-making, sectoral mechanisms for dynamic permission autonomy, technical standardization for agent identity and interconnection, voluntary task-execution standards, and industry initiatives on agent conduct and trusted assessment. Building on these practices, the contribution proposes a consumer-autonomy framework covering agent and responsible-entity transparency, task- and risk-based authorization, preservation of basic services and accessible human alternatives, execution interruption and human takeover, and traceable conduct with unified redress. It further suggests that these mechanisms may be implemented progressively according to national regulatory capacity and market maturity, with particular attention to developing countries and persons with specific needs. The meeting is invited to consider the proposed framework as input to the work of Question 5/1 on emerging consumer-protection issues and proactive regulatory approaches.
|
Word